Background:The diagnosis for smear-negative pulmonary tuberculosis (TB) is very difficult.Proteomic finger-printing of sera is a potentially useful tool.Methods:This study analyzed the results of the proteomic fingerprinting in sera obtained from active TB patients and controls using the surface-enhanced laser desorption ionization time of flight mass spectrometry (SELD1-TOF-MS) and protein-chip technology.The peaks were detected and analyzed,and a diagnostic system was developed.The protein peak was identified using high performance liquid chromatography (HPLC)-tandem matrix-assisted laser desorption/ionization-TOF-MS (MALDI-TOF-MS).Results:Around 50 protein peaks were found significantly different between the TB patients and the controls (P<0.01).Three protein peaks at m/z 5643,4486 and 4360 were selected for system classification and the development of a decision model.The model distinguished the TB patients from the controls with a sensitivity of 96.9% and a specificity of 97.8%,respectively.The diagnostic accuracy was up to 97.3%.The one most discrepant protein peak at m/z 5643 seen in sera of active TB patients was identified as orosomucoid.Conclusions:A diagnostic system for active TB was developed using mass spectrometry and protein chip technology and required only small-volume serum samples.One potential protein biomarker at m/z 5643 was identified as orosomucoid.